PE: Do you see the FDA placing any restrictions on the use of AI and machine learning over time? What would prompt such action?
Fakhouri: As I mentioned in my keynote interview, I get asked: Will the FDA regulate large language models? Are we going to ban the use of generative AI? My answer is generally not to regulate linear regression. We want to examine the data and information produced by modeling techniques and ensure that the information is reliable. So we're not trying to prohibit or ban certain AI or machine learning type algorithms. What we're really interested in is how robust, accurate, and reliable the information from these models is.
PE: What do you think the future of AI and machine learning in pharmaceutical R&D will be in the short and long term?
Fakhoury: We are very excited about the use of AI. You will see AI increasing efficiency in many parts of the drug development process. Thinking outside of our usual focus, such as discovery and protein folding, can shave years off development time. This is very exciting. Because it could bring safe and effective medicines to market sooner. It can also fill certain gaps in rare diseases. For example, there is great potential to leverage AI to accelerate drug development. I think in this situation, it's going to be a long-term thing. In the short term, I think we're all on this adoption curve, whether it's industry or academia with regulators. You need to train your staff, bring in the right expertise, and develop the right tools to solve the right problems. That will take time. Therefore, I believe that the near-term use of AI will be primarily about low-hanging fruit, such as improved operational efficiency, which will lead to faster development of safe and effective medicines.
